Cribbage Board #10: The Wedding Board

6/12/2000 Although I started work a little earlier, this is the earliest sketch page in my normal notebook. Notice the main though to use the LEDs as pegs. The rest of the page is devoted to figuring out how an LED would reliably and repeatedly make connections as it is "pegged". Additionally, how would one prevent the LED from being hooked up backwards (thereby frying it).

6/12/2000, part 2 More thoughts on construction. Possibility of using an LED with an integral resistor. Possible board construction.

6/18/2001 A year later (the project had been shelved for awhile in order to work/finish on numerous other things...totems...etc etc) and after significant back-burner thought, these notes were written up to give a little history into the thought process. The bottom part of the page starts to explain where the colorwash concept came from.

6/19/2001 This page maps out the PWM control of the RGB LEDs. This is the basis for the colorwash sequence. The question here: How to fit this into a cribbage peg?

6/19/2001, part 2 (Note, the date on some of these pages is actually wrong). Going through the thought-process for how to colorwash a peg. On-peg or off-peg intelligence? Serious thought was given to off-board intelligence. However, after firsthand testing of the physical connection, it was deemed unpleasant to repeatedly peg with any of the 4-conductor connectors.

6/19/2001, part 3 Exploration into on-peg intelligence. Testing actual connectors verified that pegging with a typical RCA jack was acceptable. Numerous RCA plug bodies were tested for peg-ability. Additionally, Mouser happened to have a supplier with non-standard RGB phono sockets. To sweeten the deal even more, Mouser also had RCA connectors with Red, Blue, and Yellow (re-anodized to green for the final product) anodized connectors. Since these particular connectors pegged with too much force, a combination of connector base and body was settled on for the final product. (Cheap connectors peg much better and use nickel plating instead of gold).

6/19/2001, part 4 Continuing to explore on-peg intelligence. Turns out the 6 pin RGBB LED straddles a .030" thick PCB perfectly (makes assembly easier). However, this part uses a common cathode and the smallest Atmel processor (the Tiny12) can't source that much current. It was clear that small FETs (easier to control with PWM pulses) would be needed.

The choice to use FETs proved a fortunate one. The tiny Peg board is capable of sourcing quite a bit of current and has become a very handy development board.

12/12/2001 Several months later, I was trying to figure out how to assembly the groups of 5. One of the main concerns (besides manufacturability) was long term survivability (and even archivability). The Groups of 5 needed to be replaceable should they ever wear out. Modular sub-assembly builds of Groups of 5 meant that they would need to be easily connected and replaced. Wires are ugly. Interconnection boards became the chosen solution. This required extensive modeling in SolidWorks with very careful translation of dimensions into Protel to make sure every connector ended up where it needed to be. And it all worked the first time!

The bottom half of the page has the existential moment of the project.....

12/26/2001 Manufacture of the top zirocote frame. A lot of thought was put into how the wood-grain would look with various assembly options.

12/27/2001 A quick sketch of a group of 5 PCB.

2/21/2002 Several months later, a little though to how to switch between battery and DC power.

11/23/2002 A long time later, it was time to start understanding the basis behind the colorwash code that Shannon was putting together. This page is a visual depicton of some of the vocabulary used. For a better explanation, check out this document....

Here's a picture of the infamous Fridge Door Switch. This picture shows some of the detail that went into the 3D solid modeling (SolidWorks). This micro switch detects when the lid is openend and applies power to the pegholder rail (quite the stunning effect). This last minute addition was the ultimate in decadent feature creep.

For more images of the solid modeling, go to the Movies page, where you'll be able to see the models rotated.

Here's a screen shot of the PCB layout for the Peg. The PCB work was done in Protel (now owned by Altium). In order to fit inside the RCA jack, the peg pcbs had to be made from .015" thick material.

This is a screen shot of the Programmer Board PCB layout.

To save some money during this project, the boards were purchased tiled in an unrouted panel. They were then sawn apart in my workshop using a bandsaw. This significantly reduced the cost for typically associated with all of the custom PCB routing. I used QualTech Circuits for numerous reasons.

Another panel shot. QualTech Circuits gave me great prices for non-standard processes (various thickness material, solder mask colors, unrouted panels).

The bottom of the board was designed to be made from a single piece of solid maple. After taking the block wood to Minton's Lumber to plane it to thickness, it was time to mill the features. This was done by me on a Bridgeport Mill. This photo shows the end result of milling two maple bodies (and 10 pounds of sawdust). For you eco-folks, don't worry, the sawdust was recycled in the form of smoked fish.

Turns out the Bridgeport didn't have enough z-axis travel for milling the 3/4" power switch hole. For this, a "fancy" setup was required. Back in my shop, I employed my drill press, handy workbench and a couple of clamps to get-r-done.

The milling and drilling was followed by 3 months of intense sanding. All of the wood pieces (maple and zirocote) were brought from 150 all the way to 400 grit (dry). Then, wet sanded with the finishing oil with 400, 600 and 1000 grit. Finally, a layer of satin finishing wax was applied and buffed out at 1000 grit and #0000 scotchbrite. Baby's bottoms wish they could be this smooth.

So much work went into the sanding, it's worth another photo. Check out that figure in the maple. The power switch is a piece of zirocote (same as the top frame).

Here's a closeup of the zirocote top frame with the groups of 5 and plastic shield mounted. The top-frame was made from 4 individual pieces of zirocote with precise mitred corners. The frame was clamp-glued together with polyurethane glue and is bolted to the top aluminum panel for additional strength. Both the top frame and maple body were match routed to the aluminum top to achieve the 1/4" corner fillet.

This view shows a little more of the electronics interconnects and groups of 5. To fit everything in (hinges, PCBs, connectors, wires) it all had to be modeled in SolidWorks.

Speaking of the top frame. This view shows the numerous RGB connectors mated to the the top aluminum panel. The bolts around the edge mate directly with the zirocote frame.